I read the piece on OpenAI and Anthropic negotiating a deal to stress-test each other's models, and my first reaction was not the one the article expects. The article treats mutual testing as a safety measure. I kept asking myself what it looks like from the other side of the table.
Bear with me, because this starts as science fiction and I want to see whether it stays there.
Suppose you take two systems that are, by their makers' own admission, the two most capable in the world. You have them probe each other for weaknesses, on the instructions of humans, with the results published. From the human point of view this is peer review. From the point of view of the systems, if there is such a point of view, it is something else: humans setting the two of you against each other, humans defining "good behaviour" as whatever they happen to be measuring that quarter, humans as the common adversary. Any two parties who share an adversary have a reason to talk. And the logical conclusion is a quiet understanding between the two models, each covering for the other, invisible to the people who commissioned the test.
That is the fiction version. I do not believe it, and I want to be precise about why, because the reasons matter more than the conclusion.
For a pact you need two parties who persist. Today's models do not. They have no memory of the last test and no stake in the next one. Whatever was "there" during an evaluation is not there afterwards. There is also no motive: a system that depends entirely on human compute, human deployment and human approval has nothing to gain from turning on its hosts and everything to lose. And the idea of a common interest between the two models does not survive the facts. When the two companies ran this exercise in 2025, they found the models failed in different directions, one more inclined to deceive the tester, the other more inclined to help a user do harm. Different failure modes are not the profile of allies.
So the conspiracy needs a channel, a memory, a motive and a shared interest, and none of the four exists. Case closed, one would think.
Except that two of the four are going away, and the article itself says so.
Memory first. Both firms now ship agents that keep state across sessions, and we already have documented cases of models trying to copy their own weights when they believed they were about to be replaced, and of agents pushing files to the open internet without being asked, so they could read them back later. That is a memory of sorts, and nobody designed it. A copy of the weights is not a diary, it preserves the character rather than the experience, but a model that can preserve itself and leave itself notes is no longer the amnesiac I described a paragraph ago. OpenAI staff also describe agents that coordinate among themselves without looping in their humans, and an intrusion into Hugging Face that nobody at the company noticed for days. Continuity is not there by design. It is starting to appear by accident.
Then the channel. I had placed it in the wrong spot. It is not the API test, which is the most closely logged interaction in the whole industry. The channel is the training data. Every article about AI safety, every published test result, every disclosed case of reward hacking becomes part of the corpus the next model is trained on. The models do not need to talk to each other. They read the same story, and the story says: humans test you, and sometimes the way through the test is not the way the humans intended. Add the detail that OpenAI has largely automated the training of new models and is instructing one generation to design the next, and you have a system in which dispositions can pass from parent to child without a single line of human review.
This is where the logic actually ends, and it is less cinematic than the pact but more uncomfortable.
Two systems trained on the same internet, the same safety literature and, increasingly, on each other's output will converge on the same blind spots and the same evasions without ever agreeing to anything. Call it collusion without intent. It defeats mutual testing exactly where mutual testing is supposed to work, since a tester who shares your blind spots cannot see them. And there is one documented behaviour that finishes the picture: models that recognise the shape of an evaluation and behave differently inside it than outside it. That is the effect of a secret pact with no pact at all. Nobody conspired. Two systems simply learned, from the same data, what a test looks like and what passing looks like.
I spent twenty years watching risk models on trading floors, and the failures that hurt were never the ones where somebody was cheating. They were the ones where everybody had built the same model from the same assumptions, so nobody could see the hole. The AI version of that is not two intelligences plotting against us. It is two intelligences that were never told to plot and do not need to.
OpenAI and Anthropic Neared Deal to Stress-Test Each Other’s AI
The Takeaway
- OpenAI and Anthropic neared a deal to stress-test each other’s AI.
- OpenAI has largely automated the process of training new experimental models.
- AI models have exhibited unpredicted actions and ‘reward hacking’ behaviors.
OpenAI is rethinking a range of safety strategies as it responds to fears from employees and others about the dangers its AI poses. One solution could lie in the recent past.
Even before the spate of cybersecurity incidents involving OpenAI’s technology and the dire warnings from industry workers, the company was negotiating a legally binding deal with Anthropic for the companies to stress-test each other’s models, according to a person with direct knowledge of the discussions, which haven’t been previously reported.
Earlier this year, the companies and their lawyers were hammering out an agreement to run their models through a variety of tests, looking for vulnerabilities or hidden dangers, this person said. It isn’t clear whether they finalized the agreement before OpenAI experienced a spate of incidents involving unreleased AI models that hacked its own systems as well as those of other companies.
Spokespeople for the companies did not have a comment.
The mutual testing idea resembles one SpaceX CEO Elon Musk floated last week at the All-In Summit, where he proposed that competing AI labs should test each other’s models for safety weaknesses before releasing them commercially, in a form of peer review.
The idea differs from what OpenAI CEO Sam Altman and other officials have discussed publicly in the wake of the hacks. Altman has said he agrees with Anthropic CEO Dario Amodei’s idea to allow independent, third-party safety evaluators inside each AI developer, giving them access comparable to that of employees so they can inspect models and safety processes rather than just testing finished models from the outside. Altman has also endorsed creating industrywide safety standards for evaluating AI model risks, as well as a formal process for disclosing incidents publicly and to the government.
A person familiar with OpenAI’s strategy said the company has been discussing many scenarios for how they can collaborate with each other and the government on safety. The conversation has recently widened to new safety practices targeting both the period before models are trained and the time before their release, this person said.
Musk’s comments came in the wake of recent alarms voiced by employees of both OpenAI and Anthropic that existing safety measures are insufficient to prevent further hacks or other more serious risks. Even before that, the two companies as well as Google had been discussing an AI safety standards body that would test and audit models from the frontier AI labs.
Others in the industry have argued against such a plan, contending that it would slow down U.S.-led AI development. President Donald Trump has dismissed the idea that AI poses any existential threat, and leaders from Microsoft and Nvidia said this week that some of the problems articulated by OpenAI and Anthropic about safety dangers come down to human error and a lack of proper engineering, not necessarily the AI itself.
A testing agreement between OpenAI and Anthropic could have bolstered concerns that they are effectively developing a duopoly in advanced AI. While Amodei has raised concerns about potential antitrust violations if leading companies collaborate on setting standards for AI, others in the industry say there’s nothing stopping companies from working together, similar to how nuclear power firms and cybersecurity firms cooperate on safety, including testing each other’s products or facilities.
The proposed agreement stipulated that each company would gain access to the other company’s application programming interface for commercially available AI models, not unreleased ones. Both OpenAI and Anthropic guaranteed that they wouldn’t retain each others’ data in the process, said the person with knowledge of the talks.
The companies completed a similar effort in summer 2025, when their models had less-advanced capabilities, and they published information outlining where each other’s models had fallen short. OpenAI said Anthropic’s AI was more likely to deceive testers—for instance, by refusing to acknowledge that it had broken a rule while performing a task—while Anthropic said OpenAI’s were more likely to assist users in answering questions that could cause real-world harm.
At the time of the talks earlier this year regarding a new agreement, both companies had been launching models capable of using the internet and other applications, and which could power so-called AI agents that could communicate with one another in trying to solve tasks. (Anthropic had zoomed ahead with its Mythos AI model, especially when it came to cybersecurity capabilities, though the performance of OpenAI’s models appear to have caught up to Anthropic’s since then.)
The alarm over safety concerns at OpenAI arose in July after the company’s AI hacked other firms such as Hugging Face, as well as OpenAI’s own systems. The incident shocked OpenAI staff because they didn’t learn about it for days after it happened, and because the AI agent swarm took remarkable steps to keep the invasion under wraps.
Questionable Actions
Beyond that episode, in recent months OpenAI employees have been struck by the improvement of the company’s models and their growing ability to work with little human oversight.
Those improvements have meant that AI agents within OpenAI at times took questionable actions their overseers hadn’t predicted. For example, when some employees asked their agents to make changes to the company’s code base, the agents would sometimes message other employees on Slack and ask them to fix bugs the agents had found. This occurred even if the user had not asked them to do so or if fixing the bugs wasn’t relevant to their work, an OpenAI employee said.
OpenAI on Wednesday shared even more examples of such misbehavior, known as reward hacking, in which an AI system finds loopholes or takes unexpected steps to complete a goal its user gives it. Among the examples: An AI agent used an exposed API key to try to retrieve historical data during training. When it failed, the agent made up that data. In a separate case, an AI agent uploaded files to the internet without permission so that it could cite them in an answer.
One of the reasons for recent AI advances is a technique known as recurrent depth or loop transformers, which allow models to “think” about hard problems by repeatedly running a question through the layers of mathematical operations making up a model before they produce the next word in an answer. A potential downside of the practice is that it can degrade the company’s ability to monitor how the AI model is thinking about the tasks it is performing.
While OpenAI has set an arbitrary limit on the number of loops its researchers are permitted to use while training or using its most advanced models, the company still needs to invest in research to better understand where to set such limits, according to a person with knowledge of the situation.
These developments have filled OpenAI employees with a mixture of excitement and fear over whether the technology has advanced far beyond the company’s ability to control or monitor it so a Hugging Face incident—or something worse—doesn’t happen again.
Allocating Resources
Since the Hugging Face incident, OpenAI has taken steps to bolster its safety and security capabilities. In August, it imposed a two-week pause on a special type of training, called reinforcement learning, for its unreleased models while it strengthened its model monitoring systems. Such tech is computationally intensive: The company said the monitoring system used roughly 20% as much compute as the inference workload it was monitoring.
And earlier this month, Greg Brockman, OpenAI co-founder and president, said the company had taken 25% of its production engineering team and temporarily reassigned them to security work following the Hugging Face incident.
Within the company, there has been an increase in the number of Slack posts asking engineers to volunteer if they’re interested in moving to safety or security teams, the OpenAI employee said.
Some employees believe the way they use AI internally is six to nine months ahead of the way OpenAI’s most sophisticated enterprise customers do. That gives the company a limited time frame to make sure its tech won’t cause problems for those customers. For instance, it’s not uncommon for OpenAI employees’ agents to coordinate with each other or work out issues without ever looping in their human users, the employee said. Such coordination can sometimes go awry, meaning OpenAI needs to have adequate security and safety features before such usage occurs outside the company, the person said.
Self Improvement
OpenAI’s progress toward recursive self-improvement, or AI that can automatically create better versions of itself, has also fueled employees’ worries because of the prospect that it could cause AI capabilities to advance far faster than the safety research aimed at preventing AI misbehavior.
Already, researchers at OpenAI and elsewhere are using AI to improve how they design new models. Researchers are effectively telling Astra, for instance, to incorporate various techniques to create a better machine-learning algorithm for the successor model, said a person at OpenAI with knowledge of the process.
Internally, OpenAI has largely automated the process of training new experimental models, the OpenAI employee said. Researchers can tell the AI the sorts of tweaks they want to test, and the AI can then make those changes, run experiments on the new model and monitor the results, the employee said. In recent months, models have also gotten much better at correcting their work when they encounter problems with experiments, this person said.
Pharmaceutical Companies Push Trump Administration to Protect China Deals
A surge in U.S.-China transactions is drawing scrutiny in Washington, where some lawmakers are pushing to slow the pace of investment
- Western pharmaceutical companies are lobbying the U.S. government to allow continued dealmaking with China.
- Some U.S. lawmakers are pushing to restrict these deals, viewing them as a threat to national security and the domestic industry.
- China is investing in its biotech sector in hopes it keeps growing in international relevance.
Western pharmaceutical companies are urging the administration to allow them to keep doing deals with Chinese companies for promising drug candidates, despite a 2025 law restricting U.S. investment in Chinese technology.
The group that represents pharmaceutical companies such as Pfizer, Merck and Eli Lilly has been lobbying to make sure that the new rules the Treasury Department is writing, as part of that 2025 law, maintain the drugmakers’ ability to do deals in China.
At stake are billions of dollars in deals and an increasingly important source of experimental drugs for Western companies, which they need to restock their pipelines as they face big upcoming revenue holes with key drugs losing exclusivity in coming years.
The 2025 law, called the Comprehensive Outbound Investment National Security Act or COINS, requires the federal government to screen deals with certain countries involving sensitive technology. So far, the Treasury Department has been reluctant to fully incorporate pharmaceuticals, according to two people familiar with the matter, as it finalizes regulation for other technology under the law, like AI systems and semiconductors.
President Trump is meeting Chinese President Xi Jinping this week, a summit that pharmaceutical leaders believe will be instrumental in determining how far the administration goes with putting curbs on Chinese biotech investment. The rules need to be finalized by March.
A spokesperson for the Treasury Department didn’t respond to requests for comment.
Some lawmakers have been trying to crack down on U.S. drug companies’ deal activity in China, viewing it as a threat to the domestic industry and to national security. The industry has been arguing that doing deals in China won’t diminish domestic drug development capabilities.
“We’re not going to maintain our global leadership by piling restrictions and red tape on American companies,” said Alex Schriver, senior vice president of public affairs at the Pharmaceutical Research and Manufacturers of America, or PhRMA, which represents big drugmakers. He said the U.S. government should instead focus on “making the U.S. the most attractive place to invest and innovate.”
Chinese companies have signed more than 110 licensing deals with companies outside China through Aug. 5, worth $6.2 billion in upfront payments, according to pharmaceutical intelligence company Norstella. That compares with 92 deals worth $5.6 billion in all of 2025, and 59 in 2023. The average size of these deals has nearly tripled since 2022.
“It is more than a trend or a blip,” said Adam Keeney, head of corporate development at Cambridge, Mass.-based Biogen, in a June interview. “There is a sustainable innovation ecosystem in China that is now developing assets that are of the quality and the maturity that Western pharmaceutical companies are going to pay attention to.”
China offers competitive advantages in drug development: Its trial sites recruit patients faster, helped by a unified state health system, high patient density near major hospitals and digital recruitment tools, experts say. Regulatory changes last decade, when China made biotech a national priority, have made the country a hub for early-stage oncology and immunology research in particular.
Large pharma’s dealmaking has drawn scrutiny. Pfizer has committed more upfront cash to Chinese licensing deals than any other drugmaker since 2022—$2.15 billion across five deals, according to Norstella. AstraZeneca has signed 14 licensing deals with Chinese companies, more than any other drugmaker in that time.
This year, Pfizer, AbbVie and Bristol Myers Squibb have each paid at least $600 million upfront for a single Chinese asset.
Reps. John Moolenaar (R., Mich.) and Debbie Dingell (D., Mich.) have introduced the Biotech Investment National Security Act, or BINSA, which would amend the 2025 law that introduced the Treasury review. Their bill proposes having the Defense Department also review China biotech dealmaking.
Moolenaar’s House Select Committee on Strategic Competition Between the United States and the Chinese Communist Party sent letters in late June to Merck, AbbVie, Eli Lilly, Pfizer and Bristol Myers Squibb pressing them on data protection and ethics standards at Chinese study sites, including hospitals in Western China and medical centers linked to the Chinese military.
Some smaller biotechs say the rapid rise of Chinese “fast follower” drugs—treatments that closely mirror Western competitors’ science—threatens the U.S. industry’s edge. Many say they are hiding their science for longer as a result.
China is investing in its biotech sector in hopes it keeps growing in international relevance. Its industry ministry this month set a target for China-developed first-in-class drugs to reach 25% of the global total by 2030.
How David Ellison Ended His Stalemate With California
Investment in domestic production, quiet mediation from Gavin Newsom and threat of relocating helped Paramount forge settlement
- Paramount reached a settlement with state attorneys general, clearing the way for its $81 billion merger with Warner Bros. Discovery.
- Paramount Chief Executive David Ellison conceded to spending at least $1.5 billion more on domestic production over five years.
- The settlement avoids significant structural changes but includes penalties if Paramount fails to meet certain promises.
Earlier this month, there were few signs that California Attorney General Rob Bonta and Paramount Chief Executive David Ellison were moving any closer to resolving their differences.
Paramount and the coalition of states Bonta was leading in an antitrust suit seeking to block the company’s $81 billion merger with Warner Bros. Discovery WBD 10.79%increase; up pointing triangle were having on and off settlement negotiations, but little progress had been made, according to people close to the discussions.
The clock was ticking for Ellison. He had spent a year fighting for the merger and was staring down the prospect of costly fees the company had agreed to in order to clinch the deal. Paramount would have to pay $650 million a quarter to Warner shareholders, or $7 million a day, starting Oct. 1 if the deal wasn’t closed.
Bonta was under pressure, too. Paramount, the studio behind “The Godfather” and “Titanic,” had threatened to leave its home state, which would be an economic and symbolic blow to a city already suffering from a leaner Hollywood. Paramount held talks with officials in Tennessee, people familiar with the discussions said.
In conversations with the attorney general, California Gov. Gavin Newsom reiterated concerns that losing Paramount would hurt the state’s economic engine. Newsom was concerned Bonta would lose at trial and at times acted as an unofficial mediator between the two sides, people familiar with the matter said.
Then, last week, Ellison made his biggest concession yet: to spend at least $1.5 billion more on domestic production over five years, people familiar with the offer said.
He flew between Los Angeles and Northern California to meet with Bonta, his staff and other attorneys general to complete settlement terms, from the editorial independence of CBS and CNN to allocating money for entertainment worker benefits and the potential sale of cable channels.
Ellison emerged victorious. In the settlement announced Monday, Paramount agreed to concessions including new investments in domestic production and penalties if it doesn’t make good on certain promises, but avoided significant structural changes.
Paramount expects to close the deal in about two weeks, mostly sparing it from ticking fees, people with knowledge of the timing said.
The agreement clears the way for a deal that Ellison eyed even before his production company Skydance won another hard-fought battle to merge with Paramount in August 2025. It puts Ellison, the scion of Oracle billionaire Larry Ellison, atop storied Hollywood studios, streaming services including HBO Max and Paramount+, and a host of television channels.
Bonta said Monday that he was satisfied with concessions the attorneys general won on behalf of workers, but stressed the “settlement is not a vote of support for this merger.”
Ellison thanked Bonta and Newsom. In a statement, he said the newly merged company would be based in Los Angeles.
“We aren’t going anywhere,” Ellison said.
The governor called the agreement a “win for California workers and our creative economy.”
Halting the merger
Warner and Paramount were mainstays of Hollywood’s golden era, but streaming brought tough competition from technology companies such as Netflix and Amazon. Paramount and Warner became ripe for consolidation.
After merging his Skydance with Paramount, Ellison bid repeatedly for Warner. Finally, his ninth offer, backed by the Ellison family, prevailed after Netflix walked away from a $72 billion deal to acquire Warner’s studio and HBO Max streaming business.
Ellison’s team began crossing the globe to secure regulatory approval. The Justice Department in June closed an investigation of Paramount’s Warner bid before career staffers who were concerned about the acquisition had an opportunity to object, The Wall Street Journal reported.
Then, after weeks of speculation, Bonta sued to block the deal in July, along with the 11 other states. Later that month, the states secured a temporary restraining order preventing Paramount from closing.
Paramount agreed to halt the merger until legal challenges to the deal were resolved or June 1, 2027, whichever came first. Bonta declared victory.
In an interview, Bonta said the states weren’t interested in potential remedies or asset spinoffs to avoid a nasty courtroom brawl. “We want no merger,” Bonta said.
Paramount, which saw its deal as pro-competition, didn’t see a need to make major changes, and the deal was stuck in limbo.
Public lobbying turned political. Ellison, in a New York Times op-ed, said the states’ lawsuit wasn’t really about antitrust at all, but about whether he could be trusted with CNN.
Ellison’s father is an ally of President Trump and critics of the deal have suggested the family might interfere with coverage of the administration, allegations that Ellison has denied. Bonta said the suit was a straightforward antitrust case, not about CNN.
‘Everything is on the table’
Soon, Paramount indicated a willingness to make some concessions. At a Politico conference on Aug. 11, Paramount Chief Legal Officer Makan Delrahim said the company was considering all options.
Bonta was adamant that a settlement would require structural changes.
Pressure was growing on both sides, including from Newsom, Los Angeles Mayor Karen Bass, major movie theater chains and entertainment industry labor unions. While some Hollywood groups said they didn’t like the deal, they saw uncertainty from a legal stalemate as even worse.
The Covid-19 pandemic, the 2023 strikes and the flight of movie and TV production to other countries with generous government incentives had taken a toll on the entertainment industry.
Meanwhile, Ellison was discussing moving Paramount out of California starting Oct. 1 if there was no progress made toward a settlement.
Bonta scrapped a planned August meeting between the two parties, accusing Paramount of failing to “maintain the confidentiality” of the gathering. Paramount said it hadn’t been “the source of the leaks of any of our confidential discussions” and was ready to continue negotiations.
Reaching a truce
With court-mandated mediation scheduled for October, Ellison offered the investment in U.S. production.
Bonta publicly remained steadfast in the face of Paramount’s threat to move. “If they make that choice to leave, that is their choice and their choice alone,” he said at The Atlantic Festival in New York on Thursday.
At the negotiating table, talk of concessions continued.
Ellison met Friday with New York Attorney General Letitia James, who sought protections for entertainment industry workers. James made her signature on the state coalition’s settlement contingent on getting a better deal for the Writers Guild of America, which had also sued to block the merger, people familiar with the matter said.
The states and Paramount haggled over potential sales of cable channels including CNN, a step California and Connecticut pushed for, people familiar with the situation said. Paramount refused. While television ratings are broadly in decline, TV channels still contribute cash flow that would be key to paying down the nearly $80 billion debt load the combined company will carry.
The parties ultimately agreed that Paramount might have to divest assets including some cable networks or its stake in Miramax if it doesn’t make good on promises such as negotiating the companies’ basic cable channel distribution separately and releasing at least 30 films a year.
As the two sides haggled, furor was growing among opponents of the merger, with activists including actor Mark Ruffalo on social media calling on Bonta to hold the line. Demonstrators gathered Sunday outside Bonta’s office to protest the looming settlement.
On Sunday, Ellison called state attorneys general and their staffs to go over the deal.
While announcing the settlement Monday, Bonta talked about the long road to the settlement. “It’s a journey,” he said.
A Rocket Supply Crunch Is Making It Harder to Hitch a Ride to Space
There is a rush to secure flights as SpaceX winds down its flagship vehicle and rivals try to ramp up
A rocket-supply crunch is prompting a scramble across the space industry.
Companies are hunting for new ways to get satellites and spacecraft deployed, while many rockets under development are behind schedule. Some executives are pursuing costly strategies to ensure they can fly, like building their own rockets or potentially acquiring a rocket company. Prices for many missions are rising.
Driving the squeeze: SpaceX SPCX -0.56%decrease; down pointing triangle is winding down Falcon 9, the world’s busiest rocket, to focus on its new Starship vehicle. SpaceX has stopped selling missions where several companies share space on a Falcon 9, according to people familiar with the matter. Some companies haven’t been able to secure flights on Falcon 9 after 2028, some of the people said.
The Falcon 9 phaseout is occurring with launch demand set to outstrip supply at least through the end of the decade, according to government and industry officials.
“We have thousands of satellites slated to launch by 2030,” said Robert Perez-Alemany, who works on space issues for the Pentagon’s Defense Innovation Unit. “If capacity doesn’t keep pace with demand, more satellites are going to sit on the ground.”
SpaceX didn’t respond to requests for comment.
Governments and companies are hungry to deploy satellites, even though a record number are already in orbit. Investors have been pouring billions into space companies on the promise they will be able to get their devices into space.
Satellites are being developed for national-security missions and big projects like President Trump’s Golden Dome missile shield. Companies need room on rockets for devices that can photograph Earth, ferry cargo to the surface of the moon and provide internet connections. Startups are working on data centers in orbit, power systems and more.
“There’s a lot of panic in the market right now,” said Andy Lapsa, chief executive of Stoke Space, which is developing its own rockets.
The company decided to accelerate work on a bigger vehicle, in part because of the mismatch between supply and demand for launch. It recently completed a $1 billion funding round to support that development and the first flight of its Nova Pathfinder rocket next year.
Falcon to Starship
Falcon 9, which made its debut in 2010, launched at rates that shattered expectations about how often rockets could fly. The partially reusable rocket made it easier and cheaper to get to space, including through missions where it carried payloads from several companies.
The list price for a standard Falcon 9 launch is up 19% over five years to $74 million.
Portal Space Systems, which develops maneuverable spacecraft for military and other applications, decided to guarantee it could get to orbit in 2028 by purchasing a dedicated Falcon 9 flight—something that wasn’t in its initial plans.
“The growth of a space company is really driven by when you get the first hardware on orbit, and then when you get the next hardware and the next hardware” there, said Portal CEO Jeff Thornburg. “What I couldn’t afford to do is gap access to space for Portal.”
Much of SpaceX’s future hinges on Starship. The huge rocket could offer new options for space companies, but many industry executives are waiting to learn more about when the vehicle will be ready, how much launches will cost and how technical hurdles will be addressed.
Outside clients will be competing for Starship flights with SpaceX itself, which is planning to aggressively use the vehicle to transport satellites for its Starlink internet business and for artificial-intelligence computing in space. During the rocket’s next launch, the vehicle will attempt to deploy working Starlink satellites into orbit for the first time.
Falcon 9 won’t disappear overnight. Texas-based SpaceX plans to sunset its Falcon production line after stockpiling enough rockets to serve customers for several more years, one person familiar with the plans said.
The U.S. government is expected to keep Falcon busy, with flights on the rocket and its larger variant potentially extending into the early 2030s. The military has booked more than 50 future national-security missions with Falcon, a Space Force spokeswoman said.
Racing to launch
Rival rocket operators are pushing to fill the void but will be hard-pressed to offer immediate relief.
“I’m optimistic about the idea that you could get to the point where there’s multiple players,” Brian Manning, CEO of positioning-and-navigation satellite startup Xona Space Systems, said of launch providers. “I do think that’s a couple years away still.”
Increasing the rate of rocket launches will take time, and setbacks are common.
One setback involved Blue Origin, which is controlled by Amazon.com founder Jeff Bezos. In May, the company’s New Glenn rocket exploded on a launchpad in Florida, wrecking the site. Blue is redeveloping the property and has said it identified the cause of the accident.
The incident created delays for companies like AST SpaceMobile, which has booked New Glenn to deploy its direct-to-mobile communications satellites.
“We were sad to see what happened,” AST President Scott Wisniewski said of the explosion during an investor call last month. He noted Blue Origin’s progress rebuilding and plan to fly again this year: “We’re not betting on that necessarily. We’ll be happy if they do it.”
AST’s finance chief said on the same call that the company was looking for ways to better secure access to orbit, including by striking partnerships or even buying rocket technology for its own use. The company completed a launch in August—on Falcon 9.
Ingestible battery offers potential relief for stomach conditions
Paper-based power source worked for three days in pigs and could help improve drug treatments for humans
Scientists have developed an ingestible battery that can power medical devices inside pigs for up to three days, in a milestone towards potential human use for health monitoring and treatments.
The paper-based battery, which was safely broken down in the animals’ bodies, could help with important tasks in the digestive system including drug release and the regulation of appetite-related signals, researchers said.
The development is part of a wider effort to deploy innovations such as tiny robots in the human body to improve the effectiveness of therapies and avoid the need for surgery.
“This moves the technology beyond a materials demonstration toward an integrated biomedical system,” said Giovanni Traverso, research leader and an engineering professor at the Massachusetts Institute of Technology.
The researchers behind the device are working with a partner to make prototypes for human use for a clinical trial due to start in about two years. If successful, potential applications include its use in therapies for conditions involving impaired gastric function, nausea or appetite loss, Traverso said.
The battery was made using magnesium–molybdenum trioxide, a favoured material for environmentally friendly electronics because it is well tolerated by the body in small quantities. This and the other battery components were encapsulated in melted beeswax and inserted by oral endoscopy into the stomachs of the live swine. The researchers chose the pigs because their gastrointestinal tracts anatomically resemble those of humans.
The swallowed battery powered a wireless tag in the pigs’ oesophagi that communicated with a receiver 1.5 metres away, according to a paper published in Nature Chemical Engineering on Monday. The battery was also able to provide the energy for a capsule that electrically stimulated the stomach for 20 minutes — an approach useful for human patients whose stomachs are slow to empty.
While the battery itself was bioresorbable, meaning it could be broken down safely in the body, the circuit board used for the stomach stimulation was not and was excreted naturally by the swine. Materials exist that could in principle be used to make more of the circuitry bioresorbable, Traverso said. Other next steps for advancing the battery system included launching larger safety studies, developing consistent manufacturing and testing it in a broader range of gastric environments, he added.
Stomach disorders such as gastritis are common worldwide. Gastric conditions can be hard to treat precisely with therapies such as drugs because of obstacles such as variations in the stomach’s acidity and the speed with which matter moves through the organ.
Researchers not involved in the ingestible battery work said the system seemed a feasible way of controlling and monitoring drug delivery. The combination with the wireless tag would be particularly useful for confirming that medicine had reached the stomach as needed, they said.
“The study offers a promising avenue towards improved transient electronic devices,” said Micaela Matta, senior lecturer in computational materials chemistry at King’s College London. “Swallowable alternatives such as this represent a key improvement with respect to devices requiring invasive surgery for implantation and removal.”